Journal of General Virology (1993), 74, 985-994. Printedin Great Britain 985 Neurons containing latency-associated transcripts are numerous and widespread in dorsal root ganglia following footpad inoculation of mice with herpes simplex virus type 1 mutant in1814 M. S. Ecob-Prince, ~* C. M. Preston, z F. J. Rixon, z K. Hassan ~ and P. G. E. Kennedy ~ 1Department of Neurology, University of Glasgow, Institute of Neurological Sciences, Southern General Hospital, Glasgow G51 4TF and 2MRC Virology Unit, Church Street, Glasgow Gll 5JR, U.K. The herpes simplex virus type 1 (HSV-1) mutant in1814 lacks the ability to trans-activate immediate early gene transcription and enter lytic replication but it can establish and reactivate from latency. We therefore investigated the number of neurons that expressed latency-associated transcripts (LATs) in animals latently infected with in1814, the rescued revertant (1814R), or wild-type (wt) HSV-1. The percentage of LAT + neurons increased with increasing doses of each of the viruses. After inoculation of equal amounts of infectious virus many more LAT + neurons were observed in animals infected with in1814 than with 1814R or wt HSV-1. Whereas the LAT + neurons in animals infected with 1814R or wt HSV-1 were largely confined to lumbar dorsal root ganglia (DRG) L4/L5/L6 (those which innervate the lower leg), in animals infected with in1814 they were also present in DRG not directly involved with such innervation (thoracic 12 and 13, L1, L2 and L3). We concluded that the large number of LAT + neurons observed with in1814 was related to the high particle numbers in the inoculum and that spread of virus was related to limited replication as well as to the low neurovirulence of in1814. This spread was not unique to in1814 but when it occurred with more virulent viruses such as 1814R or wt HSV-1, it resulted in the death of the host. Introduction When herpes simplex virus (HSV) is inoculated intrader- mally into a mouse it establishes a latent infection in the sensory neurons of the ganglia that innervate the dermatome (Stevens & Cook, 1971). In the case of the footpad this is mainly the dorsal root ganglia (DRG) of lumbar segment 5 (L5), with some contribution also from L4 and L6 (Greene, 1959). Although latent viral DNA can be found in the relevant ganglia (Efstathiou et al., 1986; Katz et al., 1990), it cannot be detected on an individual cell basis by in situ hybridization for reasons that are not yet understood (Stevens, 1989). However, there is limited transcription of the HSV genome during latency and latency-associated transcripts (LATs) ac- cumulate in the nucleus of neurons in mice (Deatly et al., 1987; Spivack & Fraser, 1987; Stevens et al., 1987), rabbits (Rock et al., 1987; Wagner et al., 1988) and humans (Croen et al., 1987; Gordon et al., 1988; Wechsler et al., 1988) where they can be detected by in situ hybridization. The major LAT is a 2-0 kb RNA which may be an intron spliced from an 8-3 kb transcript (Farrell et al., 1991) or an alterative transcription unit using the same promoter as the 8.3 kb mRNA (Spivack et al., 1991). The function of LATs remains unclear but viruses from which the LAT gene has been deleted can establish, maintain and reactivate from latent infections (Javier et al., 1988; Ho & Mocarski, 1989; Steiner et al., 1989; Izumi et al., 1989; Block et al., 1990), although reactivation in some cases is delayed or less efficient (Leib e t al., 1989 a; Hill et al., 1990; Trousdale et al., 1991).This reduced reactivation may, however, be the result of fewer latently infected cells in the DRG from which re- activation can take place, supporting the idea that LATs have a role in the efficient establishment of latent infections (Sawtell & Thompson, 1992). Nevertheless, the presence of LATs in a cell is thought to be a marker of latent virus infection. The percentage of neurons in which a latent infection becomes established is very low. Detection of viral nucleic acid suggests that between 1% and 5 % of the neurons are involved (Galloway et al., 1979; Stevens, 1989; Speck & Simmons, 1991 ; Rock et al., 1992). How- ever, when DRG were serially sectioned to quantify LAT + cells, only 24 (0.65 %) were found in 3500 neurons of the L5 ganglion (Tenser et al., 1989). We therefore investigated whether a mutant of HSV, which was less able than wild-type (wt) HSV to enter a lytic replication cycle, would establish larger numbers of LAT + neurons. The mutant in1814 (Ace et al., 1989) has a 12 bp 0001-1350 © 1993 SGM